This guide explains what happens when you upload a document and how to read the result.
Uploading a document
Upload the document into a case, or attach it on the Overview page. Attorly accepts PDF, Word (DOCX and DOC), plain text, Excel (XLSX and XLS), PowerPoint (PPTX and PPT) and images (PNG, JPG, TIFF and BMP), up to 10 MB per file. Text-based files give the most accurate results.
Scanned documents
A scanned PDF or a photo has no text layer. Attorly reads it with optical character recognition, run by Mistral, once you have agreed to send the file for that purpose.
Which law applies
Attorly works out which law governs the document: the choice of law the document makes, else your company’s default, else your own country. If none of these is known, Attorly asks you to choose before it analyses the document.
Which model analyses it
One AI model analyses the document: the model you chose as your default, your organisation’s default, or the one Attorly picks when the choice is left on Auto. The analysis shows which model produced it.
Reading the result
The analysis has a summary, an overall risk level and a risk score from 0 to 100, the document’s clauses with a risk level each, its key terms and recommendations. The risk levels are Low, Medium, High and Critical.
Comparing versions
To see what changed between two versions of a contract, use Compare under Tools (Pro and Enterprise).